Tuesday, August 12, 2014

Why Conservation is Failing, Part 8. If it's on a map, it must be so.

Why Conservation is Failing, Part 8.  If it's on a map, it must be so.

Research has directed me to a few global data sets for conservation, like the Alliance for Zero Extinction (AZE) and the World Database of Protected Areas (WDPA).  First let me state such databases are potentially extremely useful and powerful.  But accuracy is extremely difficult and any user should be strongly cautioned of the possible shortcomings and poor quality of the data.  As I have been looking at the sites, such caveats are not obvious.  And this can lead to real problems, especially when other conservation organizations use the information, like the Integrated Biodiversity Assessment Tool (IBAT), which many organizations, government offices and businesses use in assessing potential environmental impacts.  Some examples:

The AZE is a consortium of 93 non-governmental conservation organizations, including all the big ones and a fair number of smaller organizations, from 37 countries.  With this sort of expertise, you would think they can create a pretty useful resource.  Their mission is to
 "prevent species extinctions by identifying and safeguarding the places where species evaluated to be Endangered or Critically Endangered under IUCN-World Conservation Union criteria are restricted to single remaining sites."

A fairly simple task-- take the IUCN list, find where the listed Endangered species are, then prioritize areas that are not already under protection.  This depends on three underlying assumptions: Assumption #1) knowing what species are endangered;   Assumption #2) knowing their distributions;   Assumption #3) having a reliable map of existing protected areas.  In the case of New Guinea, we lack all three, and the resulting outcome reflects the multiplying effect of combining poor data sets.

The AZE sites for PNG look pretty promising-- just eight sites.  That implies the limited range endangered species of PNG lacking protection could be fully covered by instituting protection at these eight sites.

Mt. Elimbari is listed because of the endangered frog Albericus siegfriedi.  This is the type locality for the species.  It is presumably listed as endangered because of the large human population and disturbance to Mt. Elimbari.  There are around 250 species of frogs known from PNG, a large portion of them are only known from a single or few localities.  It stretches credulity to think protection of this species on Mt. Elimbari would in any way ensure frogs are well protected.  Any species known only from a single locality is vulnerable to extinction with a single logging project or mine.  Violation of Assumption #1.  Most of PNG's frogs are listed as "data deficient" by the IUCN.

Another site is Telefomin, listed to protect the Bulmer's Fruit Bat.  But it is known from other locations in PNG, as far east as Eastern Highlands Province-- violation of Assumption #2-- We don't really know the distribution of the species.  Kemp Welch River is listed to protect the bat Pharotis imogene, but the species was recently located 120 km away.  Again, Assumption #2-- identifying priorities based on ranges we don't really know.
The Alliance for Zero Extinction map of priorities for PNG based on IUCN Endangered and Critically Endangered Species that are not in protected areas and can be protected with a single site.  To people who know PNG, this seems seems a puzzling, almost random, selection of sites that omits many endangered species and species known only from single locations...

The AZE sites have 8 mammal species covered-- implying the other 20 species of mammals in PNG listed as endangered or Critically Endangered by the IUCN are found in protected areas.   Violation of Assumption #3.  Most of these are not in protected areas, since so little of PNG is protected.  IUCN lists just two Endangered Bird species, which one might quibble with, but nonetheless neither is covered by AZE, so they are assuming these two are already under protection.  One of them, a seabird, we do not even know where it nests, so we shouldn't assume it is protected.  IUCN overall lists about 90 species as endangered in PNG, the AZE priorities cover 8 of these.  Does anyone vaguely familiar with PNG, and presumably someone is among the 97 AZE organizations, really think there is anything like this level of protection in PNG??!   

I fear the data and maps were put together by people good with GIS and making maps, admittedly doing the best with poor data, but there comes a point when drawing a line on a map based on "best available data" causes problems.  I am reminded of the old maps WWII aviators had of New Guinea with a big blank area where the mountains are with "Believed not to exceed 10,000 ft" inscribed.  Hundreds of lives were lost when planes entering cloud encountered one of the many peaks exceeding 10,000 ft in New Guinea....

One of the key sources of information for what is protected in PNG comes from the World Database of Protected Areas.  Travel to this site and you will find the base map of PNG's protected areas used by AZE and many other conservation organizations.  This database is so comprehensive, it includes just about anything ever mentioned as protected in PNG-- a full 71 protected areas.  It includes things like the Moitaka Wildlife Sanctuary Sanctuary [sic] which is a little defunct zoo outside Moresby that was down to a lone crocodile for years before it died.  It includes the Baiyer River Sanctuary, which was another zoo that closed around 1988.   Many on the list I never heard of, and most of the ones I have heard of are only protected in name-- mining, logging, hunting, you name it happens within the "protected" area without limit.  The data are an egregious violation of Assumption #3; the map of existing protected areas is worse than misleading.

So what is the message?  Sure these organizations with GIS labs are trying to help.  Certainly the data are probably better from other countries, like the United States and Canada.  But the approach that works there, and makes sense there, fails when applied to countries like Papua New Guinea. 

But everyone trusts a map.  We grow up being taught that maps are realistic representations of the Earth itself.  If you want to get from A to B, you trust the map.  When we want to learn the geography of Earth we study maps.  This is a carryover from the days when maps were drawn from reality.  Lewis and Clark walked across North America, Mason and Dixon surveyed an amazingly straight line that marks the boundary of Pennsylvania and Maryland.  But now maps can be generated from data.  No one walked the boundaries of protected areas.  Had they tried they would have discovered no such boundary exists.  No one thoroughly maps the distribution of species before they are entered in a database and there for map making.

These are part of the vast resources online for conservation planning; others are published on paper in reports and articles.  They find their way to government planning offices, they are used by people making environmental impact statements, they are waved in front of donors to extract funding.  We need maps.  But we need to put the right caveats on them, and when there are too many caveats, I'd rather we were honest and just said "we don't really know."  Better to invest on what is needed to find out, than to keep flying.  In PNG at least, we are headed for the conservation equivalent of flying into cloud, trusting our maps.